ML20125A539

Ack Receipt of Re Foreign Power Reactor Designed by GE Found Inoperable

ML20125A539
Person / Time
Site:	Monticello
Issue date:	12/09/1974
From:	Hanauer S US ATOMIC ENERGY COMMISSION (AEC)
To:	Webb R MASSACHUSETTS, UNIV. OF, AMHERST, MA
Shared Package
ML20125A542	List: ML20125A539 ML20125A559
References
NUDOCS 9212080386
Download: ML20125A539 (1)

Text

- -. -.

y r

'W C

Dr. Richard E. Webb A

Research Aseeciste d

Department of Environmental Sciences University of Massachusetts Amherst, Massachusetts 01002 Dear Dr. Webb Thank you for your letter of November 18.

I'm sorry I was out of town and only now have gotten to answer it. We foreign power reactor -

which was found to have its scram system totally inoperative was the Kahl reactor in Germany. This was a 2 design. I do not know who manufactured the actual relays, initially installed, that were-re-moved in the incident, nor do I know who manufactured ' the repiscenent relays whose manufacturing deficiency caused the incident.

It is worth noting that we had a problem of similar nature during the commissioning of one of the 3 reactors.(Monticello).

In this case I believe the relays were 3-type HFA. This makes one wonder whether the earlier incident might have been the same batch.

I i

don't kase whether it is worth asking the Germans or not. Enclosed t

you will find copies of 3 and AEC reports and an AEC " ROE" evalua-tion on the Monticello failures.

l l

It seems to me that the lesson to be learned from such incidents is not that a relays are bad, because they're not; they are an accept-able product made by a reliable manufacturer.

Like all human pro-ducts and most natural ones, they fall short of perfection. Errors in design, manufacture, installation, and operation of safety-related components are possible, have happened, and will continue to happen.

We objective is to achieve an acceptable level of safety in the presence of the known impossibility of gerfection. We do this through defense in depth, redundancy, diversity design reviews, and vigilance during operation.

Sincerely, DR 7897 Original Signed by

(

' - ~ '

~ vh r. U. ikeuer L

l Stephen H. Hanauer, Director

{

Office of Technical Advisor -

Regulation Case u

12/1/74 L.

Memo dtd 6/2/70 to J.P. 0'Re111y L-1 yd"@i,f"1.gs.

'"M',,'#h/'-

om,,

/

C " u i aEd"7/19/nrEE loro "4

-- m

._/

. e.? t t 9... s9 v

cs

...rrsca Mr... se_..

. 7.

-e 12/h.

,12/[/74 12/(/,74, om i,

12. 4 /...........

12/f/74

,/74 ron. u c-m e n.. m u cu m -

p '

.o-...,,o.,.

9212080386 741209-PDR ADOCK 05000263 P.

PDR-

. =.

..)

I l

{

s' t 4 (m,

~ s.. s t

C.

UNITCD STAT' S h

'd ATOMIC ENZH3Y COMMISS!ON

. O<.I' i

civision c r co..eu,wec 1

acasc:..i 1

m n o 23:v:.c;.tca.o

{

'M'/

Gt.EN Et.t.vN, ILLihots t0137 i

e m \\

c I

fi June 2, 1970 4

4 4

4 1

J. P. O'Reilly, Chief. Reactor Inspection and Enforce mnt stanch Division of Co.plianew, lind q u.a r t e r 1

z COMPLIA';CI INQUIRY MEMORANXM i:0RTEERN STATES POYER CO::PE'" (KO. sit 131.LC0 i

DCCKI.T No. 50-2 5; i

WS A-*/ FATL"R2 0F ru.NT PR0C"". N Es 12 U.'.*5 4

4 Ot.ri n., t raatine incpe: Loc <

19 70, the c u,;uc inspet:;"

j scs.c.f n.d that amr -

t sa: fr.Ila i : ;;.cr ;.a w. ring ne : t o nt.. and preopean

~.*ta.

eir

' u re ' ya wrc ali Lane tt.i

0. ~ 1

.CE) Type Kn 55 :e. ) ; t e : ::..

au.

!..<; ular at a d 2;

)

t.'. s

.actct precre: ton w:

O S' t r.c,rt try :a*r a n: w :

l cs

. '. sy s c u.

. :<,13,

c c:.n

~

t r 11;

' c t,. : :.."?:.t l.wie" sy.: :ns.

i

. t.

of u,.

.. a) sat

.a n..'st la power la :hese y' : r..

. m unsare :nlace i

t. orief aescription of tnt wate is 3: vet Solow:

J 2

March 2.1970: Durinr, the func:lenal checkout of the RPS, L:1 :.y 5A-K9D ' Control Valva 7 st

.c %:e strca 3ypass) fc.il ad to.iror aen -;.ressuto switch 5-14D s.:s. a::ua:c2.

Ic uas fo..r.c h:e sna. th z

..ovable con:cet bloch was giver, a cup, :he rday dre pr; o.:.

F r ec-4 then on the relay operat.;c ): 3 rty ar.:.. c :::.d :..,.- ; n.

repeat ::self.

A wiri n

.31am suen. an ta:er;-itear g r 3uac ;,.,

{

suspacted :o have caused caa ! ller:.

1

' rca Durir.L e e

. :,.. c;...;, h.c.y 3... ;4A (Law ca..m; s.

l "2cu,

. re.= failec :a c:,, e t.

.:he

c

;,,;a wi,ca 3. ;A

.m i

?:tus:ed. Again it was fot.ad that the relay re;.ac.sn i.u..

a,gx of hnad.

On:e the relay released, everythic; worked normally s

r.s>

rae,:oadicion would not repeat itself.

1 j

.:c.rch 2 3, 1970: Relay icA-Kf D (Gro.;p Three valvc iso.agion, f an 2.

ta drop our during the Reactor Water Clec.nup Syste: ?reop.

The ralay was arcppad out by hand so that the Cleanup Prcop,gould ba continued.

.--}

As in the previous failurcs, the condition saald r.oc repeat itaalf.

It uca decided.:: hat if another suspected fr.

j

.c< aho:1d occur the eircumstance surroundin;; the incident voulu bc

..diLJ in detail bef arc

ouching the relay.

i I

,..a c.J,1 V u.1.,,

t e

pu9d9 F

,,; w,,,,

ev.

...w

-=

.;-=^im;~m-W

-T %&.,,jf 5

.e >

3+.

~_.

s e

I 0

s a

t j

P. O'nailly 2-June 2, 1970 s.

i

)

1 l*

I

)

April 21, 1970: Relay SA-K3A (Mcia Steam Isolation Valvo Closure Sert:0 j

failed to drop out when IsIV 2-86A was operated. All the field wiraa j

.<ere lifted one by one and finally, the relay was removed from cha panel still in the picked up condition.

Th2 relay was then drcp;ac out, using the tension adjusting screw, disassembled, and examined for 3

sbnormal conditions, but none were Cecc.d.

i j

GE was contceted at San Jose (NIO) and Philadelph'in (Reicy Produe:

Dspertment) and inforced of the situac:on.

It was decided that the j,

CE Icadt an for the RP5 Preop would fly :: Railciciphia with two of the failed ralays A third relay vas r_e.. :n 9ta Jose.

In Pall :ctinh.. various ale::ricti t.

24;han... :As:t are par;;rc;d i

on th. f ailed relays.

Ic was (cana :ne.; :he ri.t-:11gning tabt.fe:

~

]

chipped oa one of the relays, and th:..n:ac: block was crackeC on :az other.

Uc.ca the self-aligning tabs wcc. eut off f: uns disecveced that with. good deal of hand force, che relays cculd be mat; :o 2 tex 4

because af uisalignmant.

It wa: thoulg. that misalignrant may hav been th: pecN10m. but this was no:

.c-

. ; ce I

beecuse cha fr.ilac rei.c.y thu was 2: :t.: Jite !!d not have.nr m.A t ra s.

J GL notified fie : persennal of ;.

t ra. - "in Tva.d 2ng:naartn), ae.:.

on April 2!, 15'v, rcquesting :bs: ;L.

..sys be inapected for cligncen:

i and chipped or bre an carrier je1Jes.

4 j

'.cy !, 1970: The replaced nelay 5A-K2A failed to drop out when it's 4

e-. was desnurgited-(The originsi !siled ?A-K2A relay hed'bcer, j

.. aced so that th RPS preo; could b4 continued.)

It van.evidenc

}

.. tha re.sy :a: n ::.Isalisaca, se cc. relty was lef t in the panel l

the coil. ires car.:cac und.. faceste casresentacit2 r.crivad l

.Y:.r.. Philadelphic ec study ;he

.. : f..

c e.d i t t ens.

T:u ;;.: L o ry representative re;urned to ?ni'.a_e.ic.,

1.c relay he rare x..tva tests after finding a omic.a :s:acn for tne faib ::.

i l

la i subsequent Field Engineerin;.N.co, dated.:2y 13, c2

c. J.2c fiald personnel that the first memo had incorreccir :s:riS e :ce cause to be chipped or broken contact carrier guides, but ec.a c x _, nt i,. e

n; c

Power Systems Manageman

Deparctent, Philadalp;.

.J at i

S!? ai somons: rated that the relay hangup was cue to aea; b:.ccirac..w Orm;;ura to the upper (centact end) polc~ piece.

The um. die, m;.:n: la i

a black chrouate primer which is appliec :o th2 polb piaces v a corre e F"*1 inhibiter.

Ouring some relay production cycico, the prize-ased did 4

no: fully cure at room temperatures. 'In some few rolsya;

--'.5 is:% o, f

i f

t i

..,~n n,

, - + - -

e t

4 J. ?. O'Rv111y Jun: 2, 1:70 I

.]

i

+

I curing f.lus primer buildup betuaan Ir tination: pernic c n he.::.; a :...

I to internal heat genarrted during initial extem.: (12 nrs ra or l u.;)

g j

energitation.

[

Tbc =cmo stated that heat genarctcd alt,o curca : u pel and ta : if no bonding occurs followint; the above hitial enitriant!on, it will i

$9' not occur, and that il bonding doct. o: cur, Ic will not recur ona. 6

)

arn.ature is

• freed from the upper pole ;ioco, and that this vac vortfice l

by repeated cesting of new relay.-

a.

.phs cy:les.

CZ ins..,v : n c.n.

curins must :x.:ce

..:acd in the fich Sy escrz!::ir; j

. HTA rs.'.g:, p aie ther AC or DC u. cs..e ily anac;., v 0J or d.taurai,u q for 48 hocla (providing at 1 cast a.M - n c.c c.t r a c.: r i e r c i:V re, u.r ir;;

I thac this be dono at altes c:: cept exr; it can bc pa::iuly esta:Ma.e.

l that a particular relay has been u or,;i:"d for t'c.:: raried or Ic es.::

and ha:i; subsequently dropped out u::0..

,a n decaerg.a:cioa. The et. a i

also stated that it is likely tnac :11 s;;..a past ;.;; la:dia; car, i

ei.t ablich this on normally energized relays, cnd :' cia., ac c.t11y da-c energiac:d relays are in a safe mode unen f.Lcked cp, so ctring. choir y

pole pieces can await the earliest ahu;ds,n.

.c. n m.o 4rscru :.ec -ha each nonoperational sita perform or verify the :, Love 42 'cour Jr.cc:m scion and perform adaitional site testing followla; curing in c: card:.n:e..ith the memo.

1

^

It was noted that the GE camo:=h cere c.utuc,.:c a ut::aan per2:ra, with four of the:n being; identified with spacific J. cili;ia.

Thea:

t were Tsurca,10IP, OYC, and Tarapur.

4 i

]

- 3re11minary count of tha relsys in the Monticello pi:.; id e nt i.:i cd m H.~i relays installed, rae ine 2 ur ut.s e.fermas ca;.t.. a.r e

_u. u : n.; relays of tha same type na ' bvc ~$een pro.i x in Sacc.te; suppit d panels.

4 1

...u eurtag and ecs: pre;r s: M?. no a z.f r. tu La:aa. : e..;. = s et r. ;

inspection, however, dtt:11ed inspec tion ar.d tests will se perior:.a on each Type liFA relay in the plant.

4 4

I i

0

• 4

w. we.

q I

4 1

4 4

n

,w--.-.-

,n.-

n - - -, -,, -, - -

en.-r

wg4-g

/

s.

(,$,)

J..'. O'Reilly 4

Ju.c 2, 1970 f

I j

.NSP personnel indicated that the matter would be reported to DRL.

't harold D. Thornburg C0:III:CDF Senior :losetor Inapcetor s

w cc:

E. G. Case. DRS R. S. anyd. Dat (2)

S. Leviae, Dat D..

J'r ovholt, ORL (Yr L.

h-..-ith. Jr., CO:HQ RLC 011. :,

s d

4 e

l 1

I l

b i

I tA98hpej i

. e 1

f f -

i I

i I

t l.

... ~.. _ _,, - - -. - - -

j nive *!e n

,s.

e, m' v o -

n.

j.3 p n.,;

r p.j,

g. ; p,,,...

e

-6,7,

i

@lly 29, 1970

,,. ' ;Wd g j.,

./.

h,.

,e.,. ; v v o.

1 7 kr. R. rmyd d!1 Assistant i, ire,.tur, Lice Jing

~J ' c

,q e.n l..,.

.N G v.

trashington, D. L. 205"-

, \\

4 s,.

C M. ',

/

!LP T :T.

U.Ff.1Y SYST..

T /,Y FAILURES h, y,.,,, ".'r,, ; ;.'

\\.,s, 48 l.4 % %,,./

%.;;*/

i' t

. c:es t t.22tirg M J. ser3 of your t.;:hoic 'l s cc,f f a requert uis n. O to pecuit'n iniw; a tion or. rcporW reNy failures, follo.;ing is a description of um fcilera,

.wM a tion t i the cause r ' c rrective tction takn.

I trust tais vill fulfill the re p t.

1.

Descrintico T. J J

In is - i l c.

.r I:vcl Mr !astre.:; li N O nr. :.;t (GID) uts infm ccd e

tc a v

<.y sy r.,

o 11: !.;. r 9 : e hF' fe. i kd r.0 c.,01.d.e.1 12-

~....

3.

n,x,, t.

.. (

r.s.. _.-f. ;..., <..r..w..-......\\.1,.,

u. n...

.u..

4.

. ~ -.

m.

.. o.- c, -

' i..

r.,

s '. - u' i i... t.. e.

b.a.

L i b.,

i

,a w...,m.. 1'....

,6

$ n.

.iy..

i.,, r..'. O.i.

• n. ^.,. o. :.

a' > r -v, ' '

c' l...' n u,a c' i.'. e.,..1..> ', "..

m.....,

m.

Resi &hi rnret' r ::ed by a DC.11: uts '.;usp;ctrJ as, a possible failure

. a c t s ? i m. Lin....

.., M it ctr.1 L. hwl.

.T ).

. 1. 4... tier M.. C..."

t.i tc ? r21:.;> ' ;

ret, 9.eJ to M t.- W.* ice fer d.cr ':'.'3mionticr..

The un6r c i ra.m...- c ; t O r' ' - s c.c R 2 lc.i pl'.J. tic cone.t t c crier v:ra chipp:i aaj 3

t,

.n r

.,3. a s r ' '

o ;i tic-c ilcre c.echanist w'.3 0.0 F: niselictr ant rc:,e i ? i r 7,' '

' r (:i

cc. ) r.9 c b 0 n g pre.,; 0. ly g'J i,.,;.. All sitas

• -c'i i

~ U.; t c..e Q in3 M.

j rr.';.0. v.iy [i l,' li.d Ca rric e ChiW 3.

At tIl d s t'.'. t i r !',

4 t4 3 V W'.'.

initi.'..

dO3 #',n ir*Tr \\'CNnt !]rCrr;rl tC fit;'d 3 Stm0r m!ida r

/tcriel

d u r.

i;a shippi. &c.p by imprond p hac ; t.g.

.4 Cu.:t in ;

testit 2: :ID end the 'anbr reoaled thet the cause of s tichit.5 ras nec tS J uri a ;....s, bu t ra '.h. -

the psi.W. eq Un pale picca U.iich achm f i.c i he L' 'Ol SPM t'..?0 pli.tC O f td0 GN0rgiZ63 relhy.'

IO pr0VG t h i s f,12Ch,-Ini m, 29 AC nir.:

r.d r v

lay hed 0.2 paint recoves flr Ua role deco and cn-plLCOU

0 Em:

'ICn 00Tcifig, hient'/-nine (2T) ~.F'2io.Gi AC l'Oliy: u n.!

a 32 llc 0 4 ** Wrh O ', % r0CeiVJd frC;l th0 V20tiOt'.

^ll 122 r3hy5 Wrd U3Uhr.C'.I i n p.7 '. 'l i C i 'i l O r t C

^12 in Y iCh th*'. fop 3r[0J.i Iniluft 3 GCOU. r?J cnd w.cc er.ar.-

., s

y.. I,,.,..

. 1 u.s, 3,,..

..,s.

.3.,,..1.,.

O.#

4 Il I a" b r S, C '.' e ". *.. ' y" i.'.*. '.. a" l'..'

r.

iv I

b is O...'T..'".'.s'.

prop ;r C,'.iratiGn,

. ? D 'J. i i a.

t.. n.:

'*S0r.Or WM perf0n.3 np C Sh ? : W,(c;t u!1ng cj... g,,

t,.

y i,,. 'a.a.,

,,,,s.,

...c,

)

.-.,a

p..

..,}

[g e..,

.,u t.. t.

t. y..

" i UO P C i ?.,'/ 3 ' ; 21'."-

.,! :n j 10 h!.d l0n appliOd to thu aim /.:ra.

...s e

i Yf A

n j(,

NN7 f

i Jul 1MO

]

id.R.DuJ Pu.

4

]

Th2 rault: of the tens are shc.m in Tcble ! (ver.aur realts in pm. thesis)

~

ead support the conclusien that the mechanism of failure is cdharone.c of the i

paint on the pale piece to the anrcture plate uhile the relcy is cr?rgized.

The fact th8t the DC relays did not. stick was u.cught to be due to their having a loeter operating terparature (no eddy current lesses).

3 i

1' TABLE I 3

i Reicy Type Quantity llu.tber Stuck i

DC

$2 0

l DC-Treated 32 0

j AC 29(5) 7(4) l AC-Treated 29(10) 0(0) i i

The seven in werativ: ralen ere manually opened ccd re-erergized 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> to datermine if h::,t curig the paint risht be a possibla cerb tgainst recurrance of the failum. All sevM n,rmtW rnmectly, nt thi erd c" 12 hour: (as did the othec 312).

On the buis of thne results una t,e recorxt.cndation of tne i

vendor all rc:ctor sites with i!IO

.pplied relays o! tha scme type cr.d vendor U.:re instructed to energize the reltys for 43 hours4.976852e-4 days <br />0.0119 hours <br />7.109788e-5 weeks <br />1.63615e-5 months <br /> unius it'could be posi-twcly estel'lished tMt a particult.c relay 1.ad been e.urgized for that period,

or longsr.

In the lctte.- case, the rclay was ac.npt frul ths 43 hour4.976852e-4 days <br />0.0119 hours <br />7.109788e-5 weeks <br />1.63615e-5 months <br /> run if it could further be proven to have operated properly upon being de-energized.

In cddition, c11 i!ID-equipped sitos uere instructed to de-erergize th_o nor= ally tsergized rel.ys cnce eve ~y 24 hcurs for a ueek follculog the 48 hour5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> run.

A l

similar instructicn t:as issued to ul0 llanufacturing for reicys in the shop.

In Jun?, word '.as recciwl frcm one of the overseas sites that relays were still l

sticking e f ter fiva _ cycie: cf operacico.

III.

Cccrective A: tion Discussicrs betueu M10 acd tne ven3r resulted in the cecclusion that the paint was acst likely than : plastic and ll.e only_ solutions tc the probica were to either rerove De p: int fecm the pela picca er to coat the armature with a 4

natorial to bhich the paint would not adhere.

The vendor recormanded a teflon p; int which h:-c teen tested and proved satisfactory. Chea requested to provida p cof thn the teflen paint ucula last for the life ci d.e relay, the vendor de-murred.

Tha natarit.1.-as too new to have real tim,e life data and the relay. life was too log ta elicN Arrhanius curves to be generated since the necessary extra-polttien wapid nat be defensible, rurth]r investim ti e 1, the verAcr revealed th2t relcys mde by others used the

[

ne p:.la piec m ariel, u m.'A, rat ;ill-C-U12, and had not ~ experienced any cerrosioa,Miicul,ies.

Also, cnat.or rolty type ir Ja by the vendor of the problem relay was ning tha spe pola picca a a2 rial uncoated.

l, i \\-

i.

J

.c.

p',

j v

f. ', J'.0 l

1 l'r' R. D':yd 49 4-l l

s 0, t:n imis af L in?t.

' ia, ':!D cn li.c '.c.r.kr dec&d ta rr.c; ee tb paint frca the (;

? th: p:Jic pietet for relays in tha ficid, in i:lD's i

shop and in the v.

c's shcp.

The vender uill leevo the pole face ur. painted for future produci

.i. A Ficid i:cro was issued by !;ID to all MID-applied sites instructing in the paisii c. oval process, precauticus to ba observed an f I

tests to be ru<i following r m t-Sly.

I Tha vsnJoe reported on July 1,1970 on tM re:ults of a corrosion test on 6 1

relcy in u:iith a pele ? ec3 uit'1 an un:cated face ues sub.r.erged in alt water i

for a p.:riod Icng cr.cegh to build up a corrosion le3 er. The pole pitc0 uns 3

reasseobled into its reley ard tests ucro run which gave no indication of j

degredation of perforance.

TI.s v nar.it:s regeested to devise end run c test 4

which unald pron that uner..~ov:-d peint from surfaces other than the face of j

the pole pie:c uculd not dy nte to the fcce end cause stiching.

It is felt j

that since only a mli p:. :M:p2 of relays with paint on the pole fece ; tech in the field and in the teus durcribed above then it most likely require' en optimum thicknats to provide sufficient adheren:e. The small at.ount which 2

might nigrcte to the face is highly unlikely to be su:h cn amount and, further, J

if the paint hus suf ficient.trength to bt.d it most likely is too viscous to p:sults of these tc,ts uill be naue available to you as soon es they are

run, i

avail;ble.

1 IV.

Conclusinns i

The reiry in quastian has b&o e.Ta tuced and used for 30 years vithout prob-j less end \\; th a rs old of hig:. reliu :lity.

For this reason, it'evs chosen for this critical cpplicatica.

Dise.ssitat uith the vandor indicate that prior tc.

filD's purci, _2 of the rei;ys which failed, the vendor moved his mnufacturing pknt an: uu. in t;:2 prec.;s c? taini,, n i pur:onn:1.

This change in person-nel and lack of control of paint thickne :. resulted in the failures.

I ical the st2p3 raken :: descri::?d abwa in pordgraph III, plus increasta testing and curv? illa:e by i:ID Ouality Contr:1 edquat:ly precluda recurran:o of this failure.

If'I ccn be of furth a assisi.e.nca in this matter, please lot me hno.t.

i Sines"cly, Q. -

i.

~

p j

I. F. STUMT, !4 nager i

Licensing Unit

.,M/C 032 IFS:jb

"+

cci S. Levy A. P. Dray J. [3arn!.rd j

l H. Hendon 1,

~,

, - ~,. - - - -

t

,N

6i PUE 71 16 ISO / '.Ii ' T

.EU,Y FM LUKES IN hi'/ DIM PE0!ECilVE SYSTEMS l

i Eu.mnry Nine instances 'o'f REU.Y FAILURES IS Tile MONITORISG AND REACTOR PEO-TECTIVE SYSTEMS OF REACTORS are described in this report. Although the significance of any individual relay failure depends on the particular circ umstances, the proper operation of most cf the control, protective, cnd engineered safety feature systems fcr nuclear power plants depends on the successful functioning of relayr.

Ee}ay fcilures cf the types illustrated by the occurrences reported here could, under different cir-cumstances, have significant advern ef f e cts er p} ant safety.

Ncne of these occurrences resulted in da age to the rea tcr nor in radiation ex-posure of any individual.

Circumrtances A.

During a period of about two months during preoperational test-ing of a power reactor, there were four failures of relays in the Reacter Protection System or the Centainment Isolation System.

In each instance, the relays that had been derigned to open when de-energized failed to drop out on loss of power., These relays were all of the same general type, and supplied by the same manufacturer.

There are approximately 400 of them in this plant. After investigation, the manufacturer directed that all relays be inspected for misalignment and chipped or broken self-aligning tabs, which were thought to be the cause of the malfunctions.

A few days after the memorandum calling for this inspection was issued, a new relay, which had been installed to replace one of those mentioned in the previous paragraph, failed to drop out when its coil was de-energized.

It was evident that this relay was not misa11gned, and since there was no obvious reason for the failure, it was returned to the factory for more extensive tests.

Further investigation disclosed that the relay malfunction was due to heat bonding of the armature to the upper (contact end) pole piece.

The bonding agent was a black chromate primer which had been applied to the pole pieces as a corrosion inhibitor. Evidently, during production of some of these relays, the primer had not been fully cured. As a re-I sult of this lack of curing and the buildup of primer between laminations, the internal heat generatien during initial extended energizing of the relays in the field for periods of approximately 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> caused the arma-

=..

ture to bond to the pole piece.

Field instructions were issued stating that the chromate primer was to be mechanically removed from the pole piece faces.

The field memorandum i

se

i

~ ~. - - -

p.-

i 4

I, 162 I

i l

ROE 71 16 also specified additional testing of the relays in their finally installed i

l condition, following removal of the primer. The memorandum was sent not only to the operator of the reactor involved in this instance, but also to operators of three other reactors that might have received similar re-lays from the manufacturer.

i B.

During preventive maintenance on the plant protective system at i

a research reactor, one of the scram relays was replaced because its con-i i

tacts appeared to be dirty and pitted.

During subsequent tests before reactor operation, when the manual scram button on the console was de-pressed, the control rods did not drop.

The trouble was traced to the newly installed relay.

An internal short circuit was found that shunted the console scram buttons, thus I

nullifying the manual scram action. All relays at this f acility had been tested on a relay checker prior to being placed into service, but the checking instrument did not include provision for testing for internal short-circuits.

All new relays are now checked for this defect when they are received from a supplier.

i l

C.

While manually withdrawing a control rod at a power reactor, the operator observed that the rod continued to move out for several seconds after the selector switch was moved to another rod. The same occurrence was observed several times during tests of rod insertion and withdrawal.

Current is supplied to the rod drive through the contacts of a relay in the rod control circuit.

The armature spring that opens the relay contacts was found to be i

weak.

Tightening of the spring resolved the problem.

D.

During a routine shutdown of a research reactor, all the rods were being driven in by operation of a gang switch.

When the rods were about six inches from full insertion, shim rod #1 suddenly stopped moving.

The operator, thinking that the rod had possibly jammed, turned the gang switch to the neutral position, and the stopped rod started mov!ng out.

The op-l erator then returned the gang switch to the insert position and shim rod l

F1 stopped moving outward but still did not insert.

By rapidly turning the gang switch from neutral to insert, the operator was able to accomplish full insertion of all the rods, including shim rod #1, Subsequent inves-tigation showed that a loose armature spring en a, relay was the cause of this situation.

h,.-m l

At this facility, the direction of rotation of each of the rod drive mctors depends upon which of two relays in the rod drive circuit is ener-gized.

The armature spring of the withdraw relay of shim rod #1 drive cir-suit had come loose, thus perritting the relay contacts to remain closed.

With the gang switch in'the insert position the rod drive motor stopped I

I

163 p

iti: 71 16 l

becaure it had tVt simultanceut 5 f enair. ene te iraert a: ? the other t-withdraw When the f ang t wi t ch was turned t o it s neut ral prsit.<n, ie-moving the i r.F e r t sicnal, the rod was frce to withdraw.

Evidently, vr.e:

the operctor rtild]> turned the rang switch f rom neutral to :nsert, hi enet gi ct d the instri relay and the ellitticn of its tentacte cicsine ccused the loose armtt re spring of the eithdraw relay to tcalien itself to open the withdrar rc. lay centacts and permit the rod drive ret er to energize in the intcrt cirection.

Replacement of the f ault:

hdtn rc-i lay corrected the ; rdlen.

In order to increase relaability additional relays were inserted in series with the existing control relays.

The springs en coth relays in < crit s vculd have to fall at the sare time in order to cause t t.e stre situation tc recur. Maintenance precedurte w<r. chanred se tb ic?4 r vill be repleted, ch(cked and visually inspected for pre;(r -

.ati.

i I

every six months.

E.

During the daily checkout cf a pulsed resecrch reactcr, it was observed that the transient rod did not drop censistently upcr A scran sienal. A relay ir, the control pcver circuit to the air soler.oia for t.ne j

transient rod,vas i:und to stick in the encrgized position inte rmi t t cnt:

With the relay stuck, the sc1cnoid is energized, and air press ure is ar-plied to the transient rod pisten preventing dropping of t:e transient j

rod until the power sup;ly is shut off.

It could not be deter.ined j

vhether the cause of the intermittent failure was residual negnetism in the relay armature or sticking contacts.

The relay was replaced, and the scram circuitry for the transient rod performed normally.

F.

A pulsed research reactor had been operated for about 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> at a constant power level, with the pulse rod withdrawn and the pulse red "UP" button depressed by hold-in power on its relay coil.

The pulse rod had performed properly on scram tests prior to this run.

At the conclu-sion of the run, when the manual scram bar was depressed, the "UP" button cn the pulse rod failed to release.

As a result the pulse red did not drop, although the other two rods scrarn.ed properly and shut the reactor down. The pulse rod dropped when the "DOWN" button was depressed but i

since the "UP" button was still depressed, the rod immediately withdrew upon release of the "DOWN" button.

Repeated action of the "DOWN" button produced the same results.

Finally, when the "DOWN" button was depressed for an estimated 15 to 20 seconds the "UP" button was released and the l

rod dropped.

I The "UP" button on the pulse rod is held in by a holding coil.

Cur-rent to the coil is supplied by the scram circuit and the current was ep-parently interrupted satisfactoril;, since the associated light operated correctly to indicate rod scram circuit action.

The investigators con-cluded that the sticking of the "UP" button may have been caused either l

by mechanical binding of the switch or by a residual magnetic flux in the electromagnet core due to hystercsis.

Since the button released some

164 rue 71 16 seconds after current collapse, it was judged that the secena

,,x.

,.g,,

was the cast likely. The switch and : oil assently was di%,u..

,~e, clean

J and reassenbled and no obvie:s defent was observed.

. u....,,.

ing, after reinstallation of the assenti), failcd to r;;r:d;,

previously observed. Operating procedures were issued to ins;....,

operator to turn of f the Reactor Key or Peactor Power innediato'

...,,g this situatinn recur.

A few weeks later, a similar situation did occur.

T h - a u re.

again remaved and tested for residual ragnetism. The call an; cy were observed to exhibit some ragnetist, which was eittinat.*

gaussing coll.

Operation was rcsurel, :lth the concurre m ef t-

~,,.

tor Safety Cormittee, after a replace ent switch was i st-; led testing of the old switch on a r;ack-

'est circuit faile: :c ~ p r.. : u.,

the effect following degaussing.

C.

During checkout prior to startup of a research react't.

v4, found

• Sat none of the automatic scrar circuits could be reset.

screr <a e t nei t ch u, a :.

.nt:s l

tigatien revealed that one vf the si:.

r rust ce in the closed position in rd;r to reset the ser:: sv4..n.

4n

)

open, tous intc:r -tin: the c :: ;: return f re. all si- ;f th-

.4;..

a 1

ser relays cc the power supp2y :::

t'.e rod st;ne t t.

The bf -::

e...cn was replaced and tne scram syst:n was reset.

]

During the investigation of the cause cf the ir..;;itv ta system to reset, analysis of tne circuit rev Ged that the na -eus.irm 5(

2.t sr:d ic shut d um circoit was not protecte. 'rer a single f a. ut t :n

'N' l

s. ; r e l uy p e e r s t; p i', for the si, cutn atit scran relays, k.e :: wr*

.: tt

o. -ix relay +.ust be enert :ed in order :: interru t the :;rr at P

tagnets in t~.e control rod drives (b-de-energi: int ne nernai;v the ragn?t tawer suppiv maaier 5 cran relay in t5s :ir: 11 : t2 rSe 'nce.id ~.. autonatic scrcn ralays wot.; not be enerr.:cd in the

.e-t cf a fai. rt of any con:-

=-t in the relav p:.: supply, suen a

. i' re cas N ; tne paver sup-( v;1ts;e t; drep eirni ;; ant!"

e ner-al O

losed, series-eired c ntacts in ths vital scran sus renain :.;*rd are de-energi: d.

To c:rrect

.. s the autceatic shutd:vn relav-2-..,

-undition, a new relay has been added :: the :1r:ui:, shich is en-erp ee3 by the strar relay power supply via a toltace divider netwerk.

Cn2 set of c:1. tacts n this new relay is use: to irterrupt the magnet the sc ram re1re

?

l p;wer t; the control rods in the event of a tai;.re 2*

i rearr supp1', or one of the scram reset switches.

Another set af contacts

"*'C

_ sea for r.n indicator light which indicates tr.at e.e relay is enert'

'd*

A r;-ilar pru; ;

in the reactar pr:tu:tten-systen at ancther resear:h re-

{

a:te r was reap-rted in ROE:

69 16, dated June 13, 1969.

sodium-cooled reactor a leak detector is ?ron d*d J EIIE i

h.

At a I

th; cuxiliary critar" coolant syster pump it sodiunleakageisdetect'f' 1

Dur.g the performance of conthP. tests af the leak detectors, the led' i

cc - *,- trip contacts

.1 t 'r. saf ety chassis sere junpered to prevent l

d h

, pin:,of rha p unp ea ch ti r e a detector was tested, caring s

'equent 1

b 1

l

i 165 i

l l

l pt, 71 1(

ttro.*al of the j urpi :, e nt. end was Inndvttt(ntly trounded resulting in high current.

The cur. trol circuit fuse opened, was irredittely itplaced and the tetecter.ta returm d to service.

A tenth later, durinc the n,unthly leak det(-tor testing, the pu p f ailed to t rip du. a ht..: was simulated by shertinr. across the detector c utput.

In.t r t ip t.:n disclos ed that the iisi. dcttctor relt, centacts in tne pump centrel circuit had becn. tid (d tcgetter, evidently as a re-sult ei the no.ntentn:t errer the previc us = nth.

1.

The signals from the r.itek (!flucnt unsers, at a povcr reacter, t

l feed int:e indicators, equitpud with ccntt ntert 1sclation tri; contacts, i

j of the type s er.e t itt e ref errt J te as r etc r-relays.

In reter-relayr, t.e ind! cater ne(dit r:vts up s cale t.n t il i t contacts an irternal trip ccn-tact. et which point the trip is in.tiated.

The irdicator cc'ter is scalcd h

by a Epcnge rtMor psket.

In thir, inrtnr.:e, Lt.e gasket evidcativ had cre-ht!dened, dried cut and cre...t c ir4o pi eces. One stall piece of gasket t.aterial drciped in:c cne of the indicatcr needle techanisms cnd preventtd it from going up-scale (the contair. rent isolation vculd still have been initiated, if re-l quired, by redundant trip units).

Thc c-erator discovered the prob 1ct, f

during the daily trip test of the systen.

Other t.eters of the same de-sign were als: f our.d to have loose piect.s of gasket. The gaskets on all the meters were replaced with a more resilient material of a guaranteed longer life.

o e

"i A tu +f

[

t

.v,.

.., -..,.